OpenAI’s Business Hub turns ChatGPT into a business discovery channel
Business profiles and branded agents put companies directly inside ChatGPT, but teams should not hand an owned customer journey to a rented channel.
What OpenAI launched
On November 10, OpenAI launched Businesses in ChatGPT in beta. Eligible companies can claim and verify a profile, update business and product data, and see aggregated mentions, clicks, and discovery topics. Business Agents are initially limited to eligible US ecommerce and online-service businesses.
The systems consequence
ChatGPT is moving from an answer layer to a discovery and handoff layer. Its agent can respond from business-supplied information, but the conversation remains inside OpenAI and the customer leaves for the company’s site to transact. Businesses receive aggregated themes rather than individual chats.
What we would do
We would claim the profile early, keep the website and business systems canonical, instrument referrals, and test agent answers against the current catalogue and support rules. We would not bind customer identity, enquiry history, or conversion logic to this surface. It should remain a replaceable adapter.
OpenAI’s launch is not a reason to move your customer journey into ChatGPT. It is a reason to treat ChatGPT as a new rented distribution layer, similar to search or a marketplace, while keeping product truth, conversion logic, and customer records on infrastructure you control. We would claim and verify the profile, publish current structured product data, and instrument every referral, but we would not make a Business Agent the only place customers can get accurate answers. OpenAI keeps the conversation, exposes aggregated insights rather than transcripts, and controls eligibility, policy, and future commercial terms. The short-term upside is discovery; the long-term risk is dependency. For a team shipping software, the build decision is clear: make your website, APIs, and internal systems canonical, connect ChatGPT as an adapter, and shadow-test whether it sends qualified traffic before committing deeper engineering effort. If the channel works, scale the feed and agent content. If it does not, you should be able to switch it off without rewriting the customer journey.
Turn insight into a working system.
We build a canonical product-content layer, a structured company website, and an on-site AI front desk, then connect and evaluate ChatGPT as one distribution route rather than the system of record. If your team depends on AI discovery or customer enquiries, it is worth talking to us.